LASER INTERFEROMETER GRAVITATIONAL WAVE OBSERVATORY
LIGO Laboratory / LIGO Scientific Collaboration
LIGO- T060267-00-D
1/04/06
Upgrading the Input Optics for High Power Operation
UF LIGO Group, IAP Group
Distribution of this document:
LIGO Science Collaboration
This is an internal working note
of the LIGO Project.
California Institute of Technology
LIGO Project – MS 18-34
1200 E. California Blvd.
Pasadena, CA 91125
Phone (626) 395-2129
Fax (626) 304-9834
E-mail: info@ligo.caltech.edu
Massachusetts Institute of Technology
LIGO Project – NW17-161
175 Albany St
Cambridge, MA 02139
Phone (617) 253-4824
Fax (617) 253-7014
E-mail: info@ligo.mit.edu
LIGO Hanford Observatory
P.O. Box 1970
Mail Stop S9-02
Richland WA 99352
Phone 509-372-8106
Fax 509-372-8137
LIGO Livingston Observatory
P.O. Box 940
Livingston, LA 70754
Phone 225-686-3100
Fax 225-686-7189
http://www.ligo.caltech.edu/
LIGO- E030589-00-D
Abstract
Sometime during S5 or after its completion, the LIGO detectors will undergo a series of technical
improvements designed to increase their sensitivity. Among these improvements, the laser power
will be increased to deliver approximately 30 W to the PRM. At these power levels, both the
electro-optic modulators (EOMs) and the Faraday isolators (FIs) will need to be replaced.
Absorption in the current FIs and subsequent heating will lead to enhanced thermal effects: thermal
lensing, thermal birefringence, and beam steering. The current New Focus LiNbO3 modulators will
also suffer from severe thermal lensing.
We propose that the current EOMs be replaced with a UF design for Advanced LIGO based on
RTP. These modulators are similar to the current New Focus modulators in electro-optics
performance, but the RTP crystal has significantly lower absorption and therefore thermal lensing.
We also propose to replace the FIs with a new design, essentially equivalent to what is proposed for
Advanced LIGO. The new FI design minimizes thermal birefringence by using two TGG crystals
and a quartz rotator which cancel thermally induced birefringence and DKDP, a –dn/dt material to
compensate thermal lensing.
Performance data and implementation issues are discussed in this document.
Acronyms
DKDP – deuterated potassium dihydrogen phosphate
FI – Faraday isolator
FR- Faraday rotator
EOM – Electro-optic modulator
IAP – Institute of Applied Physics
MMT – suspended mode matching telescope mirror
PRM – power recycling mirror
QR – quartz rotator
REFL – symmetric port reflected beam
RTA – Rubidium titanyl arsenate
RTP – Rubidium titanyl phosphate
SM – suspended steering mirror
TFP – thin film polarizers
TGG – terbium gallium garnet
2
LIGO- E030589-00-D
1 Electro-optic modulator
1.1 RF Modulation Requirements
The requirements on optical modulation, including modulation frequencies, modulation depths, and
relative stability of the modulation frequency and amplitude1, for Advanced LIGO are described in
the IO DRD, LIGO-T020020. Since these requirements are at least as stringent as those of initial
LIGO, we adopt them here.
1.2 Modulator Material's Properties
1.2.1 Optical and Electro-optical Properties
After a literature survey, consultation with Crystal Associates and Raicol Corporation, and
experiments in our lab, we have chosen rubidium titanyl phosphate (RbTiOPO4 or RTP) for the
modulator material in advanced LIGO. A related material, rubidium titanyl arsenate (RTA), also
meets requirements and we are carrying that as an alternate. The standard modulator material, used
in initial LIGO, lithium niobate (LiNb03), is not satisfactory from the point of view of thermal
lensing, damage, and residual absorption. A summary of the properties of RTP, RTA, and LiNb03
is shown in Table 1.
Table 1 - Material Parameters for Selected EOM Crystalsa)
Properties
Units/conditions
RTP
RTA
LiNbO3
Damage Threshold
MW/cm2,
(10ns, 1064 nm)
>600
(AR coated)
400
280
nx
1064nm
1.742
1.811
2.23
ny
1064nm
1.751
1.815
2.23
nz
1064nm
1.820
1.890
2.16
cm-1 (1064 nm)
< 0.0005
< 0.005
< 0.005
r33
pm/V
39.6
40.5
30.8
r23
pm/V
17.1
17.5
8.6
r13
pm/V
12.5
13.5
8.6
r42
pm/V
?
?
28
r51
pm/V
?
?
28
r22
pm/V
Absorption coefficient 
1
3.4
LIGO T020021, “Sideband Requirements in Advanced LIGO, Part I”, G. Mueller, et al.
3
LIGO- E030589-00-D
nz3 r33
pm/V
239
273
500 kHz, 22 oC
30
19
Conductivity, z
cm-1, 10 MHz
~10-9
3x10-7
Loss Tangent, dz
500 kHz, 22 oC
1.18
-
Dielectric const., z
a)
306
Data from Raicol, Crystal Associates, Coretech, Note that the reported data are not all consistent. Moreover,
many are strongly temperature and frequency dependent, particularly the conductivity and loss tangent..
The largest EO-coefficient is r33. The optimum configuration is propagation direction in the
y-direction; applied electrical field and polarization of the light field in the z-direction. The
modulation depth is then:
  m 
U
L
r33 n z3 z

d
where L is the length of the crystal, Uz the voltage across the crystal, and d the thickness in the zdirection. The modulation depth of m = 0.5 leads to:
Uz
L
 620 V for RTA
d
and
= 708V for RTP.
For typical dimensions, L = 2.0 cm and d = 0.4 cm, Uz = 124 V. A resonant circuit, with the RTP
crystal as the capacitor, reduces the required input voltage by the Q of the resonator, typically 5-20.
The low electric conductivity (high resistivity) of RTP reduces the electric power
consumption inside the crystal to well below the microwatt range, negligible compared to the heat
generated by the resistance of the leads and the electrodes. Moreover, the heat produced by the
laser beam itself is more critical because of its nonuniform spatial distribution.
1.2.2 Thermal Properties
The thermal properties of RTP and RTA, measured by several workers, are summarized in Table 2.
Thermal lensing scales as the parameter Q:
Q
dn 
dT 
Table 2 - Comparison of Thermal Properties of RTP, KTP, and LiNbO3
Properties Units
RTP
RTA
KTP
LiNb03
dnx/dT
10-6/K
-
-
11 (b)
5.4 (c)
dny/dT
10-6/K
2.79 (a)
5.66 (a)
13 (b)
5.4 (c)
dnz/dT
10-6/K
9.24 (a)
11.0 (a)
16 (b)
37.9 (c)
4
LIGO- E030589-00-D
x
W/Km
3
(e)
2
5.6 (d)
y
W/Km
3
(e)
3
5.6 (d)
z
W/Km
3
(e)
3
5.6 (d)

cm-1
Qx
1/W
-
-
2.2
4.8
Qy
1/W
0.047
0.94
2.2
4.8
Qz
1/W
0.15
1.83
2.7
34
< 0.0005 (e) < 0.005 (e) < 0.005 (e) < 0.05
a)
"Temperature-dependent dispersion relations for RbTiOPO4 and RbTiOAsO4," I. Yutsis, B. Kirshner, and A. Arie,
Appl. Phys. B 79, 77 (2004). The x-axis values were not measured, but should be close to those along the y axis.
b)
Handbook of Optics, no wavelength specified,
c)
Crystal Technology, Inc.
d)
only one value given, no axis specified.
e)
For the KTP, KTA, and RTP absorption coefficients, the largest value is for electric field along z, as that is the
direction of highest dc conductivity and RF loss tangent.
1.2.3 Comparison with LiNbO3
The larger optical absorption in LiNb03 makes for much stronger thermal lensing than in RTP or
RTA. Moreover, the 7:1 anisotropy in dn/dT results in a large ellipticity in the beam profile as
observed in some of our experiments. In addition, surface damage was also occasionally observed
at > 30 W powers. Thus, LiNb03 will not suffice as a modulator material for Advanced LIGO. It
also will not suffice for use in an upgrade to 30 W.
1.2.4 RTP
Comparing RTP to LiNb03, the ratio of dn/dT to for the z-axis in LiNb03 and for the worst case
in KTP is similar. The smaller absorption coefficient is where the benefit of RTP is found.
The expected thermal lens in the modulator can be estimated from the sag change over the
beam profile:2
s  0.1
LP dn

L
P
 0.9 m  2

 dT
[10 / cm] [1.5cm] [200 W ]
2
Justin D. Mansell, Joseph Hennawi, Eric Gustafson, Martin Fejer, Robert L. Byer, David Clubley, S. Yoshida, and D.
H. Reitze, “Evaluating the effect of transmissive optic thermal lensing on laser beam quality using a Shack-Hartmann
wavefront sensor”, Appl. Opt. 40, 366-374 (2001).
5
LIGO- E030589-00-D
This leads to a thermal lens of
2
 w   [ m] 
w2
 
f 
 1m
.
s
 [mm]   s 
For an absorption coefficient of xcm-1 and a 1 mm beam waist, the sag change is
50 nm and the focal length of the induced lens would be 20 m. Raicol reports a total absorption
coefficient for RTP as low as 50 ppm/cm (including scattering and 2nd harmonic generation), one
order of magnitude lower than the absorption coefficient used above. The subsequent sag change in
the beam would be about 0.4% of the optical wavelength for 200 W power, the thermally induced
lens in the 200 m range, and the higher-order mode content less than a percent.
Note that the absorption coefficients given in Table 1 include also non-thermal losses, such
as scattering, 2nd harmonic generation, etc. Therefore, the absorption values are upper limits for the
thermal problems we could expect and it is possible that heat producing absorption coefficient will
be lower. According to information from a variety of sources, the heat producing absorption
coefficient in KTP is much lower than in LiNb03.
1.2.5 RTP thermal lensing measurements
We have made a number of studies of thermal lensing in RTP. One example is given in the IO
Conceptual Design Document (LIGO T-020022). Fig 1 illustrates the measured beam divergence
angle for various input powers up to 60 W. For these measurements, the RTP crystal was located
at z=32 cm, and measurements were taken with a Photon, Inc. Beam Scan. Measurements were
also taken with and without the RTP in place to determine the background thermal lensing in other
RTP
Fig 1 – Measurement of the change in divergence angle as a function of
incident laser power. The RTP crystal was located at 32 cm along the Zaxis. Changes were recorded as function of power at 1 W, 30 W, and 60 W
with and without the RTP in place.
6
LIGO- E030589-00-D
optical components. The upstream optical components exhibited significant thermal lensing,
perhaps due to absorption in the AR coatings, and thus the spot size in the crystal decreased as the
power was increased: 700 m (1 W), 550 m (30 W), and 450 m (60 W). A positive thermal lens
in the RTP, as expected for + dn/dT, would shift the waist closer to the crystal position and increase
the divergence angle.
From Fig. 1, we see no difference in the waist position or beam divergence at 1 W and 30 W. At
60 W, a very slight decrease is seen in divergence angle, although there is no noticeable shift in the
waist position. The measurement error associated with the fitting was greater than the observed
change in the beam divergence angle, thus within the measurement accuracy there is no thermal
lensing observable.
1.3 Modulator Design
We have made RTP-based phase modulators using 1.5 cm long x 0.4 cm width x 0.4 cm thickness
crystals. The crystal electrodes were gold over titanium. Clipping losses of a w = 360 m Gaussian
beam radius (roughly the value in the current EOMs) are below 60 ppm.
1.3.1 Mechanical Layout
The modulator is enclosed in an industry-standard housing, nearly identical in details to the
modulators used in initial LIGO. The housing, which was manufactured in-house, is shown in Fig.
2. The crystal is on the right hand side of the enclosure, and the impedance matching circuit on the
left. There is an SMA connector on the side. The housing is made of aluminum.
Fig 2 – Housing for the UF-designed electro-optic modulators. It is based on New Focus housing. The
connector coming from the housing is an SMA connector.
7
LIGO- E030589-00-D
1.3.2 Impedance matching
Fig. 3. shows the equivalent circuit of the matching network. The matching circuit is designed to
have an input impedance of 50 
resonantly enhances the RF voltage at the crystal by Q,
where Q is the quality factor of the resonator.
Fig. 3. Equivalent circuit of the matching network.
We have made matching networks for f = 19.7 MHz and 180 MHz. The crystal has C = 6
pF and the loss (associated with the electrodes and the electrical wiring) is about 2 . The shunt
capacitance is in the 20 pf range. The 19.7 MHz circuit had Q = 20 while the 180 MHz circuit had
Q = 3.
1.4 Modulator Performance
1.4.1 Low Frequency
The f = 19.7 MHz modulator gave a modulation index m = 0.2 with 5 V rms drive. Based on this,
12.5 V rms will be required to give m = 0.5. The RF generator will need to provide 4 W into 50
1.4.2 High Frequency
The f = 180 MHz modulator gave m = 0.2 with 30 V rms drive. Based on this, 75 V rms will be
required to give m = 0.5. The RF generator would need to provide 120 W into 50for a
modulation index of m = 0.5
1.4.3 RFAM Measurements
Pure phase modulated light has a constant intensity. Defects in this are RFAM. RFAM is not
objectionable by itself, as the dispersive response of the interferometer and the cavities turn phase
modulation into amplitude modulation. However the requirements on fluctuations in the amplitude
modulation must meet the requirements discussed in LIGO-T020020.
We have made a number of measurements of the RFAM generated by the RTP modulator.
This is a difficult measurement, because the RFAM is near the limit of measurements. Moreover, a
power-stabilized, 180 W, single-longitudinal-mode laser is really needed for these measurements.
8
LIGO- E030589-00-D
The best results to date used a pump-probe method with a 40 W heating laser and a 200
mW NPRO probe laser. The RFAM on the probe beam was measured with a calibrated diode using
a spectrum analyzer at 19.7 MHz. The dc power of the probe was also measured. The results are
shown in Fig. 4. The upper panel shows the ratio of the power in the probe beam at 19.7 MHz to
that at dc. The lower panel shows the power from the heating beam.
FIG. 4 – Measurement of RFAM induced by absorption/heating of the pump laser. The absolute change as the
power is increased fro m 0 to 40 W is 5 x 10-5 , the RFAM swings at a given power level are ~ 1 x x 10 -5 . This could
be further reduced by applying temperature feedback to the housing.
With the heating beam off, the ratio of ac to dc powers is about 1x10-5. That the RFAM is
not zero is attributed to the alignment of the probe beam to the modulator axes. The fluctuations in
the RFAM are at the 1x10-6 level. The measurement bandwidth was 300 Hz.
The main effect that one can see as the power is stepped up, correlated steps in the RFAM. This is
clearly due to the fact that the increased power raises the temperature of the modulator and its
housing, The change in temperature degrades the alignment of the modulator to the probe beam. If
the modulator is realigned, the RFAM drops to the initial level. Because the heating laser is very
9
LIGO- E030589-00-D
noisy, the fluctuations in the RFAM increase, but there are quiet times when the fluctuations are
about the same as without the heating beam.
Further measurements are underway.
1.4.4 Damage
The intensity distribution in a Gaussian mode with total power Po is:
2
P(r )  Po
2e

r2
w2
w
The intensity in the center is therefore:
P( r  0 )  100
kW
cm 2
approximately 6000 times below the damage threshold quoted for 10 ns pulses. Although still to be
verified, it seems unlikely that this power level will damage the coatings (or crystals). Raicol
quotes a damage threshold for the uncoated RTP crystals of 1 GW/cm2.
We have subjected an RTP crystal to 90 W of 1064 nm light for 300 hours, with no damage,
induced absorption, or other changes noted. We have also destroyed a crystal instantly when the 90
W beam touched one of the electrodes.
1.5 Appendix: RTP properties, from http://www.coretech.com.cn/RTP.htm
Rubidium Titanyl Phosphate (RTiOPO4 or RTP)
Applications and Advantages:
◇an excellent crystal for Electro Optical applications
at high repetition rate.
◇High Damage Threshold,low Insertion Loss
◇not induce piezo-electric effect .
◇.Non-hygroscopic
◇High Extinction and Contrast Ratio
Standard specification:
Dimension tolerance:
Flatness:
Surface quality:
Scratch/Dig
Parallelism:
Perpendicularity:
Angle tolerance:
±0.1mm
/8
@633nm
10/5
better than 30 arc sec.
better than 30 arc min
△  < 0.5°,△  < 0.5°
10
LIGO- E030589-00-D
AR coating:
AR coatings
>90% central area
Clear aperture:
Transmitting wavefront
less than /8 @ 633nm
distortion:
Physical and Chemical Properties:
Crystal Structure
Orthorhombic,
Lattice Parameters
a = 12.96 A°,b = 10.56 A,c = 6.49A°
Density
3.6 g/cm3
Melting Point
~ 1000 °C
Ferroeletric transition
temperature
~810°C
Mohs Hardness
~5
Thermal Expansion
Coefficients,/°C
a1=1.01x10-5, a2=1.37x10-5, a3=-4.17x10-6
Hygroscopic Susceptibility No
Dielectric Constant
eff = 13.0,
Color
Colorless
Ionic conductivity
10-8 S/cm
(room temperature, 10kHz)
Optical properties:
Transmitting range:
Sellmeier equations:
ni2=Ai+Bi[1-(Ci/λ)2]-Diλ2
Absorption coefficient:
Nonlinear properties:
Nonlinear Coefficients:
350~4500 nm
index A
B
C
D
nx
2.15559 0.93307 0.20994 0.01452
ny
2.38494 0.73603 0.23891 0.01583
nz
2.27723 1.11030 0.23454 0.01995
< 0.0005 cm-1 @ 1064 nm < 0.04 cm-1 @ 532 nm
d15 = 2.0 pm/V
d24 = 3.6 pm/V
d31 = 2.0 pm/V
d32 = 3.6 pm/V
d33 = 8.3 pm/V
deff for type II SHG@1064nm= 2.39 pm/V
11
LIGO- E030589-00-D
Electro-optic
coefficients:(pm/V)
Static Half Wave Voltage
@1064nm:
Damage Threshold:10 nsec
pulse
Xcut Propagation
Y cut Propagation Low frequency
Low frequency
r13 = 10.6
r23 = 12.5
r33 = 35
r33 = 38.5
1445 V for a pair of 4x4x10mm
1,600 V for a pair of 4x4x20mm
1,700 V for a pair of 6x6x7mm
2,400 V for a pair of 6x6x20 mm
> 600 MW/cm2 @ 1064 nm,
12
LIGO- E030589-00-D
2 Faraday Isolator
2.1 Proposed Faraday Isolator Design
2.1.1 Overview and conceptual design
The Faraday isolator (FI) design proposed for the LIGO upgrade3 is based on the design that will
be used for Advanced LIGO. Figure 5 displays the schematic design of the FI. It consists of a
Faraday rotator (FR) with two terbium gallium garnet (TGG) crystals and a quartz rotator (instead
of a single TGG crystal as in conventional FRs), a ½ waveplate, a negative dn/dT material
(deuterated potassium dihydrogen phosphate, KD2PO4, or ‘DKDP’) for thermal lens compensation.
In this design both thermally-induced polarization distortions and thermal focusing are reduced
through the use of compensating elements. For depolarization compensation, the beam passes
through two 22.5° TGG-based rotators and a reciprocal 67.5° quartz rotator placed in between
them. In this configuration, the polarization distortions that the beam experiences while passing the
first rotator (through the absorption of the beam and subsequent spatially-dependent birefringence)
will be partially compensated in the second. For compensation of thermal focusing, absorption of
laser light in the negative dn/dT material defocuses the beam and compensates the thermal focusing
induced by absorption in the TGG. More details can be found in the papers by Khazanov, et al. 4,5
and Mueller.6
DKDP Thermal Lens
Compensation
Faraday
TGG Crystals
Crystal
Polarizer
QR
FR
H
/2
Polarizer
H
Figure 5 – Conceptual design of the new Faraday isolator. FR = Faraday rotator;
TGG = terbiu m galliu m garnet; QR = 67.5quartz rotator; H = magnetic field; / 2
= half waveplate.
LIGO-T050252-00-I, “Enhancements to the LIGO S5 Detectors”, P. Fritschel, R. Adhikari, and R. Weiss.
E. Khazanov, N. Andreev, A. Babin, A. Kiselev, O. Palashov, and D. H. Reitze, “Suppression of Self-Induced
Depolarization of High-Power Laser Radiation in Glass-Based Faraday Isolators, J. Opt. Soc. Am B. 17, 99-102
(2000).
5
E. Khazanov, N. Andreev, A. Mal’shakov, O. Palashov, A. Poteomkin, A. M. Sergeev, A. Shaykin, V. Zelenogorsky,
Igor Ivanov, Rupal Amin, Guido Mueller, D. B. Tanner, and D. H. Reitze, “Compensation of thermally induced modal
distortions in Faraday isolators”, IEEE J. Quant. Electron. 40, 1500-1510 (2004).
6
G. Mueller, R. Amin, D. Guagliardo, Donavan McFeron, R. Lundock, D. H. Reitze, and D. B. Tanner, “ Method for
Compensation of Thermally-Induced Modal Distortions in the Input Optics Components of Gravitational Wave
Interferometers”, Class. Quantum Grav. 19 1793–1801 (2002).
3
4
13
LIGO- E030589-00-D
2.1.2 Performance goals, design
To operate effectively at 30 W powers, we adopt the following performance goals for the new FI
design:
Table 3 – Performance specifications for the new FI
Parameter
Goal
Comment
Optical throughput (%)
> 95%
Limited by absorption in TGG and DKDP, surface
reflections in the FI components (total of 16 surfaces)
Isolation ratio (dB)
> 30 dB
Limited entirely by extinction ratio of thin film
polarizers7
Thermal lens power (m-1)
< 0.02
Leads to <2%
Thermal beam drift (rad)
< 100
Based on dynamic range of RBS actuators 8
reduction in mode-matching
Fig. 6 – Schematic drawing of the FI, showing fro m front to
back: i) 1/ 2 waveplate, ii) TPF, iii) FR, iv) 1/ 2 waveplate, v)
DKDP, and vi) TPF. The breadboard and auxiliary mirrors are
not shown. The separation distance of the initial ½ waveplate
and TFPs depends upon the specific interfero meter.
In addition, the FI must meet general requirements for in-vacuum materials set forth in LIGOE960050-B-E “LIGO Vacuum Compatible Materials List”, D. Coyne. A complete list of
requirements can be found in “Faraday Isolator Specifications for Advanced LIGO”9.
The physical layout of the FI is shown in Figure 6, and consists of the usual FI elements
(polarizers, waveplate, FR) as well as the DKDP compensating element. The clear aperture of the
system is 20 mm, the same as the FI currently in H2. The specific separation distance of the FI
components depends upon the layout of the HAMs and differs for H1 (L1) and H2.
7
As we discuss below, it may be possible to use calcite wedge polarizers (which have extinction ratios in excess of
105), which would improve the isolation specification to > 40 dB.
8
Ken Franzen and Valera Frolov, personal communication
9
LIGO-T050226-00-D, “Faraday Isolator Specifications for Advanced LIGO”, D.H. Reitze, et al.
14
LIGO- E030589-00-D
The two-crystal design of the FR mandates a larger housing and magnetic field. This has
consequences for the in-vacuum layouts. The large housing eclipses the laser beam in H1 and L1.
The FI must be moved from its current position between SM1 and MMT1 to a position in between
MMT1 and MMT2 (but still in HAM1), and SM1 will need to be repositioned. In addition, the
‘acceptance angle’ of the FI (defined by the isolation ratio sensitivity to input angle) is reduced
relative to the current FIs. For H2, it may be possible to leave the FI in its current location between
SM1 and SM2, at a reduced isolation ratio. These changes are described in more detail below.
2.1.3 Size and weight
The FR housing is 15.5 cm in diameter by 16.1 cm long. The FR weighs 30.8 lbs (37 lbs with
stand). The total FI assembly will add approximately 40 lbs to HAM1 (H1, L1) and 30 lbs to
HAM7 (H2)10 and require the removal and repositioning of some counterweights.
2.1.4 Choice of polarizers
The performance of the FI depends on the type of polarizer used; both calcite wedges and TFPs
have been tested. Both offer advantages and disadvantages which we summarize here.
2.1.4.1 Calcite wedges
Calcite wedge polarizers use the natural birefringence of calcite to separate orthogonal
polarizations. The polarizers used in these experiments were low absorption 2.5 mm thick
polarizers with a 4.3 wedge angle. They were obtained from the IAP.
2.1.4.1.1 Advantages of calcite wedge polarizers

Calcite polarizers possess very high extinction ratios (>105), and thus provide for the
optimal isolation performance. The isolation data presented below was obtained using
calcite polarizers.

Calcite polarizers offer higher overall transmission than TFPs (~ 97% vs 93% for TFPs)
2.1.4.1.2 Disadvantages of calcite wedge polarizers
10

Because the calcite polarizers rely on the (small) difference in the ordinary and
extraordinary index of refraction, the exit angle of the ordinary and extraordinary rays is
quite small, on the order of 10 mrad. In order to separate the REFL beam from the incident
beam, a propagation distance of at least 1 m is required before the pickoff mirror can be
placed. This in turn requires that the REFL beam reflect off both the MMT1 and the SM
(H1, L1) or SM1 (H2) before pickoff.

The thermal drift of the calcite polarizers as the loading power is increased is a factor of 20
worse than TFPs, but still within the 100 rad limit.

The wedge introduces an angular displacement of the beam, requiring a small change in the
layout. We do not believe that this is a significant disadvantage.
The H2 FI was replaced in 2003; the H2 FI assembly weighs more than the older H1 and L1 FIs.
15
LIGO- E030589-00-D
2.1.4.2 Thin film polarizers
The TFPs used in these experiments were obtained from CVI and LZH. Most of the experiments
were done with CVI polarizers.
2.1.4.2.1 Advantages of TFPs

The thermal steering of the TFPs is negligible at powers up to 30 W.

The rejected REFL beam comes off of the polarizer at 45, making it easy to direct the
beam to the ISC table periscope
2.1.4.2.2 Disadvantages of TFPs

The extinction ratio of the TFPs is ~1000. This places an upper limit of the FI isolation
ratio of 30 dB.

The transmission of the TFPs is lower than that of the calcite wedges.
2.2 Measured Optical Performance
We have performed a series of measurements of the proposed FI design, including optical isolation,
thermal lensing, optical throughput, and thermal beam drift.
2.2.1 Thermal birefringence and optical isolation
Optical isolation was measured using the ratio of the power transmitted through the first polarizer
to reflected power from the HR mirror: optical isolation in dB = 10 log (Pin, backward/PT, backward). For
these measurements, high quality calcite wedge polarizers were used (extinction ratio > 105) to be
limited by the thermal depolarization rather then the extinction ratio of the polarizers.
QR

Pol
P
TGG
TGG
In, forward
PT, b ackward
P R, backward
H
H
Pol
DKDP
P
HR
Mirror
T, forward
PIn , back ward
Fig. 7 – Measurement of isolation ratio for new FI design. The polarizers (Po l) are high quality calcite wedges
with ext inction rat ios > 105 . The isolation ratio is measured as 10log(PT, backwards / PIn, backwards )
Figure 8 shows the measured isolation ratio as a function of single pass power for both the new
design and the original LIGO 1 EOT FI design. At 30 W, we obtain ~ 46 dB isolation. If TFPs are
used, the expected isolation ratio would decrease to ~ 30 dB.
16
LIGO- E030589-00-D
50
45
Optical Isolation
40
Current EOT Isolator
Compensated Isolator
35
30
25
20
15
0
20
40
60
80
100
Incident Laser Power
Fig. 8 – Measured isolation ratio as function of incident laser power for the
proposed design (red circles) and current LIGO 1 FI (b lack squares).
Because the quartz rotator in the FR is birefringent and oriented such that the optical axis is
parallel to the laser propagation, the isolation ratio depends sensitively on the angle that the
propagating beam makes with the optical axis of the FR. Slight angular deviations of the
propagation axis of the laser w.r.t. the quartz optical axis produce linear birefringence whose
magnitude depends upon the relative deviation angle of the beam with respect to the polarization
axis of the beam. The worst case occurs for deviations 45 w.r.t to the polarization axis. For this
case, the measured sensitivity to angle is shown in Fig. 9. The isolation falls off steeply for
incident angles greater than  0.25 ( 4.4 mrad). The acceptance angle of the FI is determined by
the apertures of the mirrors bracketing the FI and their separation distance. For H1 and L1, the FI
will be placed downstream of MMT1 (in between MMT1 and MMT2); this defines an acceptance
angle 4k =  (aperture of MMT2)/(distance from FI to MMT2) ~ 0.0375/15 =  2.5 mrad. For H2,
the FI will be placed in between SM1 and SM2, corresponding to an acceptance angle 2k =  32
mrad. For alignments in which the H2 beam is substantially displaced from the centers of SM1 and
SM2, reductions in isolation to below 30 dB may result. Thus, it may be prudent to reposition the
FI to downstream of the MMT1 in H2.
17
LIGO- E030589-00-D
Isolation ratio (dB)
42
40
38
36
34
32
30
28
perp. or par. to laser polarization
45 deg to laser polarization
26
-20
-15
-10
-5
0
5
10
15
20
Incident Angle (mrad)
Fig. 9 – Isolation ratio as a function of incident angle w.r.t the optical
axis of the FR. Black points correspond to angular deviations parallel
or perpendicular to the laser polarization axis and represent the
minimu m depolarization. Red points correspond to angular deviations
at 45to the polarization axis and represent the worst case.
2.2.2 Thermal lensing
Thermal lensing was determined by measuring the beam divergence as a function of incident laser
power for i) the FI alone (without the DKDP), ii) the DKDP alone, and iii) the entire thermally
compensated FI.
DKD
P
Pol
TGG
QR TGG
/2
Pol
WinCAM
z
Fig. 10 – Measurement of the thermal lensing in the FI. In this case, the DKPD was placed before the FI and
only the single pass lensing was measured. However, the location of th e DKPD in the upgrade will either be in
between the polarizers or between the FI and the PRM
18
LIGO- E030589-00-D
0.20
Rotator only
Compensator only
Rotator and Compensator
-1
Focal Power (m )
0.15
0.10
0.05
0.00
-0.05
-0.10
-0.15
0
20
40
60
80
100
Incident Power (W)
Fig. 11 – Thermal focal power of the FI as a function of incident laser power. The
black squares show the focal power (in m-1 ) for the FI without thermal co mpensation.
The red circles display the measured focal power of the DKDP. The green triangles
display the measured focal power for the fully co mpensated FI.
Figure 11 displays the thermal lensing induced in the FI as function of single pass power. For these
measurements, there was no backward propagation. The total thermal lensing (green triangles) at
30 W of incident power is negligible; at 70 W, a focal power of approximately 0.025 m -1
(corresponding to a thermal focal length f = 40 m). The magnitude of thermal compensation
depends on the absorption coefficient and the length of the DKDP crystal. It is difficult to control
the absolute absorption in DKPD since it depends sensitively on the stoichiometry of deuterium.
For these measurements, it was known in advance from absorption measurements that the DKDP
would overcompensate for the thermal lens in the FR. Better results can be obtained by further
thinning of the DKPD crystal; the DKDP will be optimized for the FIs used in the LIGO upgrade.
2.2.3 Thermally induced beam steering
The measurement set-up for determining the thermally-induced tilt as the power is increased is
shown in Fig. 12. Measurements were made for an incident power of 30 W; zero time corresponds
to the time at which power was sent into the FI. These measurements were performed without the
DKDP crystal.
19
LIGO- E030589-00-D
QPD
ND Filter
Wedge
Wedge
TGG
QR
TGG
/2
Pol
Pol
HR Mirror
Fig. 12 – Experimental set-up for measuring power-dependent thermal steering.
2.2.3.1 Calcite polarizers
The thermal drift with calcite wedges as polarizers for 30 W incident power is shown in Fig 13.
Measurements at lower powers show that the magnitude of the drift is linear in power. Thus, a
thermal drift coefficient of 1.3 rad/W is inferred. The vertical and horizontal drifts are
comparable in magnitude; the origin of the vertical drift is unknown at this time.
50
Drift (rad)
40
30
20
10
Horizontal
Vertical
0
-10
-20
0
1000
2000
3000
4000
5000
Time (s)
Fig. 13 – Change in angle of the rejected FI beam versus time for calcite
polarizers from a cold start (no power) to 30 W.
2.2.3.2 Fused Silica Thin Film Polarizers
Thermal drift using TPFs with 30 W incident power is shown in Fig 14, corresponding to a thermal
drift coefficient of 130 nrad/W. The vertical drift is slightly larger than the horizontal drift, but
both are quite small when compared to the calcite wedges.
20
Drift (rad)
LIGO- E030589-00-D
3.0
2.5
2.0
1.5
1.0
0.5
0.0
-0.5
-1.0
-1.5
-2.0
-2.5
-3.0
-3.5
-4.0
Horizontal
Vertical
0
1000
2000
3000
4000
5000
Time (s)
Fig. 14 – Change in angle of the rejected FI beam versus time for TFPs from a
cold start (no power) to 30 W.
2.2.4 Optical throughput
The optical throughput was measured for the full FI using both calcite and wedge polarizers. The
fractional transmitted and rejected light are as follows:

FI with TFPs:
o Percent transmitted: 93.3%  0.3%
o Percent rejected: 90.3%  0.3% (defined as Pr, backward/Pin, backward, where Pin, backward
is the light power incident on the FI in the reverse direction and Pr, backward is the
rejected light power)

FI with calcite wedge polarizers:
o Percent transmitted: 98.3%  0.3%
o Percent rejected: 94.6%  1.0%
2.3 Modifications to the current in-vacuum optical layout
2.3.1 4K IFO
Figures 15 and 16 show the new layout for the repositioned FI for the 4km IFOs. In addition to
repositioning of the FI, other changes to be made include:
a) In order to accommodate the large size of the FR housing, both SM1 and MMT1 will need to
be moved from their current location towards the edge of HAM1. SM1 will be shifted ~ xx cm
along the MC beam propagation direction, and MMT1 will need to be shifted approximately xx
cm toward the edge of the table. In addition, the angle of incidence of the beam on MMT1 will
increase from 4.9 to 10.8.
21
LIGO- E030589-00-D
b) The half waveplate currently located between MC3 and SM1 which rotates the polarization
from ‘s’ to ‘p’, will be moved to after SM1. This will decrease the transmission of the SM 1
from ~2% to ~0.2%. In addition, the reflectivity of SM1 depends on the angle of incidence,
which will further decrease the transmission. The total transmitted power of SM1 for 30 W
incident is expected to somewhere in the range 30-60 mW, be we need to measure it to be sure.
c) The transmitted SM1 beam used to PSL intensity stabilization will be re-routed to avoid the
new FI. This will require the addition of two fixed kinematic mounts and mirrors
d) The breadboard of the FI (not shown in the drawings) assembly will be wedged to account for
the beam rise from MMT1 to MMT2. (HAM 2 is lower than HAM 1 so the global coordinate
axes slope downwards at both sites. The slopes are different, so site-specific breadboards are
required.)
e) Beams dumps will be moved to their new appropriate positions.
Figure 15 – Proposed optical layout for 4K HAM 1. Clamps for the fixed
mirrors will be moved to the appropriate location for tie down. The FR
housing does not clip the beam from MMT3 to the PRM since their
elevations are different.
22
LIGO- E030589-00-D
Figure 16 – Close-up of 4K Faraday isolator location.
23
LIGO- E030589-00-D
2.3.2 2K IFO layout
Figure 17 shows the new layout for the 2k IFO. This replacement is relatively simple. Only the
transmitted SM1 beams and reflected RM beam lines will need to be moved.
Should the isolation ratio dependence on angular alignment through the FI prove to be a problem, it
is possible to move the FI downstream of MMT1, although getting all of the ancillary diagnostic
beams (not including REFL) may require complicated paths.
Figure 17 – Proposed optical layout for 2K HAM 7.
2.4 Magnetic field coupling to IO suspended mirrors
The FR dual crystal design uses a more complicated magnetic field configuration to achieve
uniform fields in the TGG crystals. The maximum magnetic field is 1.5 T; this is slightly more than
the 20 mm aperture Electro-optics Technology FI currently in use in H2. The magnets are
composed of sintered Nd:Fe:B, as are the LIGO FI magnets.
A measurement of the magnetic fields and an analysis of the coupling of the Advanced LIGO FI to
the mode cleaner mirrors have been performed for Advanced LIGO based on the requirement that
the fields do not induce frequency noise above that required by Advanced LIGO.11
The requirements derived for the maximum magnetic field gradient to induce a dynamic force on
the mode cleaner mirrors are:
f 
6
 2 B||max

  1.35 x 10 
 10 Hz 
3/ 2
 3.89 kg  Gauss


2
 m  m
(1)
where B|| is the component of field oriented parallel to the magnetic dipole on the mirror and m is
the mass of the mirror. The requirements derived for the maximum magnetic field gradient to
11
“Analysis of Stray Magnetic Fields from the Advanced LIGO Faraday Isolator”, T060025-00-D, G. Mueller, et al
24
LIGO- E030589-00-D
induce a dynamic torque on the mode cleaner mirrors (which couples with an off-centered beam to
induce a length change) are:
Bmax

r
 f 

 3 x 10 
 1 Hz 
3/ 2
6
 8.3 x 10 3 kg m 2

I




2
 0.0107 A m
 Gauss

2
 m
(2)
where B is the component of the magnetic field perpendicular to the dipole moment  of the
mirror magnet and I is the moment of the inertia of the mirror.
These requirements were derived assuming stack motions of the Advanced LIGO seismic isolation
platforms. Using the measured ground motion at LLO, transfer functions of the current HAM
stacks, and m = 0.25 kg and I = 1 x 10-4 kg m2 for the current small optics mirror, the same
analysis yield requirements for the upgrade. At 100 Hz, we find that:
 B
2
max
|| 
Gauss
 1.35 x 10
m2
11
Bmax

r
 3 x 10 8
Gauss
m2
The measured fields and computed gradients are shown in Figure 18. The gradients are well below
the required values for distances greater 5 cm and orders of magnitude below for all suspended
optics in the layouts shown above.
25
LIGO- E030589-00-D
Fig. 18 – Measured axial magnetic fields and computed field gradients based on fitting the axial field.
2.5 Vacuum compatibility
The prototype IAP FR was not designed or assembled with LIGO vacuum requirements in mind,
and it has not yet been certified for in-vacuum operation. Some parts of the FR housing will need
to be replaced (brass, e.g.) before undergoing baking, although it will be impossible to completely
disassemble the housing and magnets for cleaning, since the orientation of the magnets in the
housing results in a strong repulsive force of the individual solenoids.
Our plan is to try a clean and bake on the FR in the next month (either at LHO or CIT) and see how
it fairs. Should it fail, we will institute clean assembly procedures (components are first clean and
baked and then assembled according to LIGO spec) for the new FIs.
2.6 Assembly, Alignment, and Vacuum Incursion Plan
1. Upon receipt, all components will be baked and checked for vacuum compliance according
to LIGO Vacuum Protocols
2. After baking, components will be assembled, aligned, and characterized in the LLO/LHO
clean optics lab. This includes:
a. Assembly and testing of the polarizers after baking
b. Assembly and testing of the FR after baking
c. Build-up and alignment of the FI assembly
26
LIGO- E030589-00-D
d. Measurement of optical throughput, isolation ratio, polarization purity, and if
possible, thermal defocusing, depolarization and associated time constants.
3. Completed FI assembly will be prepped for transport to the LVEA.
4. Vacuum installation plan – TBD depending other planned activities.
27
LIGO- E030589-00-D
3 Costs
3.1 Electro-optic Modulator
The itemized cost per EOM is as follows:
Item
Vendor
RTP crystals
Raicol
Quantity Unit Cost
1
$2600
Total Cost
Machined Components
UF
1
$400
$400
Electronics
UF
1
$100
$100
$2600
Shipping
$20
TOTAL
$3120
3.2 Faraday Isolator
The itemized cost per FI is as follows:
Item
Vendor
20 mm Faraday rotator
IAP
Quantity Unit Cost
1
$10000
Total Cost
Half waveplates
CVI Laser
1
$530
$530
TFPs
LZH
2
$1000
$2000
(Calcite polarizers
IAP
2
$1000
$2000)
DKDP compensation plate
IAP
1
$1000
$1000
UHV pedestals and clamps
New
Inc.
Focus, 4
$28
$112
UHV Rotation Stages
CVI Laser
3
$149
$447
UHV Faraday rotator pedestal
UF
1
$20
$20
UHV
UF
1
$50
$50
UF
2
$30
$60
mirror UF
1
$30
$30
$10000
Breadboard
UHV polarizer adaptors
2” diameter
mount
fixed
Shipping
$200
TOTAL
$14449
28
LIGO- E030589-00-D
We assume that there will be a sufficient number of spare DLC mounts available in the optics labs
for re-routing the MC transmitted beam. If not, either DLC mounts can be ordered at additional
cost, or fixed mounts can be used. Also, some spare 2” supermirrors will be needed.
29